ITS Concall Minutes 20101214
ITS Teleconference - December 14, 2010
Present: Gerald Beuchalt (GB), Paul Knapp (PK), Dale Nelson (DN) Mead Walker (MW), Gunther Schadow (GS)
- Role call and agenda
Call to Order
- 4:10 pm Eastern
Datatype Schema Naming
It has been reported that the naming of Datatype Schemas which the HL7 tooling creates or includes uses the Datatype version as part of the name which ihibits simple replacement with the updated file and creates significant needless work for implementers.
Motion: That the name of the schema file for any given datatype version shall be the name of the major version of the datatype, for example datatypes 1.0 and 1.1 would use datatypes.xsd, datatypes 2, 2.03, 2.5 etc would all use the file name iso-21090-Datatypes.xsd. Any more detailed point release info would be contained within the file. Dale/Mead 4-0-0
UUID Case Constraint Backward Compatibility Issue
Gunther identified an issue with Datatypes 1.1 that uppercase is required for UUIDs in R1.1 whereas mixed case was permitted in R1.0.
Problem is that Abstract Datatypes 1.0, and 1.1, specify the case of UUIDs to be Upper, and schema for R1.1 has the constraint to upper, but the widely used schema for R1.0 does not have the correct constraint. The issue then being that either we issue a technical correction for R1.0 or relax the constraint in the R1.1 implementation to be backwards compatible with the omission in R1.0 – which is expected to be less egregious.
Motion: “The group recommends that we relax the constraint in R1.1 in the Datatypes doc and the schema for backwards compatibility, and note this for future correction.”. Dale/Mead 4-0-0
Datatypes for Publishing
HL7 has a concern about what Datatypes should or technically can be published in the ballots or normative editions - it would be easier or preferrable to have only 1.
ITS will host an open session Wed Q2 in Sydney, with invites to Publishing, IC, M&M and others to determine what Datatypes should be used for publishing and whether it is possible to then transform to Datatypes 1.1 or 1.0 for actual implementers. If R1.1 is a reasonable target then should R1.1 be taken to normative?
5:15 pm Eastern
- Sydney WGM January 10-14 2011